Introduction

The central problem of HIV-1 pathogenesis is the decline of CD4+ T lymphocytes, which precedes the progression from asymptomatic infection to the acquired immunodeficiency syndrome (AIDS). Mechanisms underlying CD4+ T cell depletion in AIDS patients still remain poorly understood.1,2 Progress in this area has been hampered by the problems in reconciling the results obtained in studies using in vitro cell systems with the events occurring under in vivo conditions and, possibly, with those observed in HIV-1 infected patients. For instance, some isolates of HIV-1 recovered from patients with CD4+ T cell depletion show no cytopathic effect on cultured human T cells, suggesting that HIV-1 infection may lead to dysfunction and eventual loss of CD4+ T cells by indirect mechanisms that do not require virus infection of the target cells.3 Recent data highly support the hypothesis that one of the major causes of CD4+ T cell depletion, both in the peripheral blood and in the lymph nodes of AIDS patients, is the virus-mediated induction of apoptosis of bystander CD4+ T cells.4,5,6,7 Of interest, various reports have shown that CD4+ memory T cells (CD45RO+) may represent the major target for both HIV-1 infection and HIV-1-induced apoptosis.8,9,10,11 However, mechanisms responsible for these phenomena are still unknown.

Human-severe combined immunodeficient (SCID) mouse xenografts represent unique and practical in vivo models to study the early events triggered by the interaction of HIV-1 with the human immune system.12,13,14,15,16,17,18 In particular, the interactions between human lymphoid cells and the SCID mouse environment occurring in these models can lead to a marked activation/differentiation of the immune cells.19,20,21 Recently, we have observed that human lymphoblastoid CD4+ T cells (CEM cells), which are exclusively permissive in vitro to syncytium-inducing (SI) T-tropic HIV-1 strains, became permissive to M-tropic strain of HIV-1 after passage into SCID mice. This in vivo acquired permissiveness was due to an increase in the expression of a functional CCR5 β-chemokine receptor and was strongly associated with the appearance of a memory phenotype (CD45RO) on CEM cells.22 In the present study, we first investigated in vivo the effects of various HIV-1 strains on the growth of CEM cells in SCID mice. Moreover, we characterized the apoptotic phenotype of CEM cells obtained from SCID mice tumors (ex vivo CEM cells) and the effects of HIV-1 infection. We found that transplantation of CEM cells in SCID mice induced apoptotic behavior in these cells and that this newly acquired property was strongly associated with the expression of the CD45RO+ phenotype. HIV-1 infection markedly increased the apoptosis of bystander CEM cells through a massive Fas-mediated autocrine suicide/fratricide. After this early phase, T-tropic strains of HIV-1 induced cytopathic effect, while M-tropic strains led to a low level of chronic infection in CEM cells, which gained resistance to spontaneous apoptosis through the development of an efficient survival program. On the whole, these data, obtained in a human-SCID mouse model, recapitulate a sequence of events that may occur in the natural history of AIDS, such as CD4+ T cell depletion and the development of latently infected cells.23,24 Thus, studies with human cells passaged in SCID mice may offer new opportunities for in vivo investigating the mechanisms underlying HIV-1 infection and CD4+ T cell depletion.

Results

Impairment of CEM cells tumor growth and enhanced apoptosis of CEM cells by in vivo acute HIV-1 infection in SCID mice

We performed an acute infection with the SF162 or the IIIB HIV-1 strains contemporaneously to the CEM cells subcutaneous (s.c.) injection in SCID mice. Consistently with a recent finding,22 CEM-SCID mouse xenografts were highly infected with both the HIV-1 strains used, while the parental CEM cells were in vitro infectable exclusively with the T-tropic strain of HIV (data not shown). This was due to a significant upregulation of a functional CCR5 on CEM cells that was crucial in rendering these cells permissive to the M-tropic strain of HIV-1, SF162.22 As shown in Figure 1A, infection with both HIV-1 strains resulted in a marked delay in the growth of CEM cell s.c. tumors in SCID mice, as compared to the uninfected tumors. To assess the nature of this virus-induced in vivo delay of CEM cell tumor growth, we compared the levels of spontaneous apoptosis, by flow cytometric analysis of TUNEL reaction, in single cell preparations of CEM cells collected from HIV-1 infected and uninfected s.c. tumors (ex vivo CEM cells), when the tumors reached 1 cm in diameter. The results showed that uninfected ex vivo CEM cells exhibited remarkable levels of spontaneous apoptosis (30–40%). However, higher levels of apoptosis (60–75%) were observed in ex vivo CEM cells from HIV-1-infected tumors (Figure 1B). Notably, histological analysis showed that both infected and uninfected CEM tumors were highly vascolarized and poorly necrotic (data not shown).

Figure 1
figure 1

Effects of in vivo HIV-1 infection on CEM cells growth and apoptosis in SCID mice. (A) CEM cell tumor growth in uninfected xenografts (□) and xenografts infected with the SF162 (▪) or IIIB (▴) HIV-1 strains; (B) Percentages of apoptotic cells in CEM cells isolated from the s.c. CEM tumors (ex vivo) uninfected ([open box]) and infected with SF162 ([filled box]) or IIIB ([hatched box]) HIV-1 strains. Points of the curves and histograms are means±S.E.M. of 15 animals in five separate experiments

Full size image

CEM cells obtained from SCID mice exhibit an apoptotic phenotype

This data prompted us to characterize the phenotype of uninfected and HIV-1-infected ex vivo CEM cells, as compared to the parental cell line. We firstly analyzed the expression of CD45RO, CD45RA, Fas, Fas-L, Bcl-2, Bcl-xL and Bax, in uninfected ex vivo and parental CEM cells. Ex vivo and parental CEM cells did not show major differences in terms of T phenotypic markers, in that both cell types were CD4+, weakly CD3+, CD8− and CD14− (data not shown). Consistently with a previous report,22 FACS analysis showed that ex vivo CEM had become CD45RO+, while parental CEM cells were CD45RO− (data not shown). FACS analysis showed a significant induction of Fas-L expression and marked decrease of Bcl-2 in ex vivo CEM cells, as compared to the parental cells (Figure 2A). The bcl2 protein level was also investigated by Western blot analysis. Consistently with the results of the flow cytometric analysis, the experiments showed significant reduction of bcl2 protein in ex vivo CEM cells as compared to the parental cells (Figure 2B). Densitometric analysis showed that there was a twofold increase of the bcl2 content in parental CEM cells with respect to ex vivo CEM cells. Consistently with the protein expression, RT–PCR analysis showed lower level of bcl2 and higher level of FasL transcripts in ex vivo CEM cells, while no differences in the expression of Bax and Bcl-xL were observed (Figure 2C). Taken together, these data indicated that the phenotype of ex vivo CEM cells markedly differed from that of parental cells. This could be the result of either a progressive differentiation of CEM cells due to the stimuli present in the mouse environment or an expansion of a very small fraction of CEM cells (expressing the CD45RO phenotype, high level of Fas-L and low level of Bcl-2) in SCID mice. This hypothesis appears unlikely in the view of the fact that CD45RO and Fas-L are not expressed in parental CEM cells.

Figure 2
figure 2

Phenotype analysis of CEM cells isolated from s.c. tumors grown in SCID mice (ex vivo) as compared to the parental CEM cells. (A) Fas-L and Bcl-2 expression in parental (dotted lines) and ex vivo (solid lines) CEM cells analyzed by flow cytometry; (B) Western blot analysis of bcl-2 protein level in parental and ex vivo CEM cells; (C) RT–PCR of parental and ex vivo CEM cells; (D) Flow cytometric analysis of Fas-L expression in uninfected (left panel) and HIV-infected (right panel) ex vivo CEM cells. All the results are representative of 15 animals in five separate experiments

Full size image

In examining the phenotype of HIV-1 infected ex vivo CEM cells, we found that the percentage of Fas-L+ cells was markedly increased as compared to the uninfected cells (Figure 2D), while Fas, Bcl-2, Bcl-xL and Bax mRNA expression did not differ between infected and uninfected ex vivo CEM cells (data not shown).

Mechanisms underlying HIV-1-induced apoptosis of ex vivo CEM cells

Then we investigated whether a Fas-mediated mechanism was involved in the in vivo HIV-1-induced apoptosis of CEM cells. We first explored the ability of both IIIB and SF162 HIV-1 strains to induce apoptosis of ex vivo CEM cells as compared to the parental cells. We observed that, at 48 h after virus challenge, both IIIB and SF162 induced a marked increase in the percentage of apoptotic cells as compared to the uninfected ex vivo CEM cells, while the parental CEM cells did not undergo apoptosis either spontaneously or after HIV-1 challenge (Figure 3A). Of interest, the values obtained with the acute infection of ex vivo CEM cells in culture condition were comparable to those observed in CEM cells collected from the s.c. tumors infected in vivo (shown in Figure 1B).

Figure 3
figure 3

Role of Fas/Fas-L interaction in the HIV-induced apoptosis of ex vivo CEM cells. (A) Induction of apoptosis by HIV-1 in vitro infection of parental (upper panel) or ex vivo (lower panel) CEM cells, analyzed by flow cytometry at 48 h postinfection. (B) Flow cytometric analysis of Fas-L expression in parental CEM cells as compared to uninfected or infected ex vivo CEM cells, at 48 h postinfection. (C) Effect of an anti-Fas blocking mAb (+αFas) on apoptosis in uninfected (CTR) (upper panel) or HIV-1 infected (HIV) (lower panel) ex vivo CEM cells, analyzed by flow cytometry at 48 h postinfection. All the results are representative of five separate experiments

Full size image

The marked upregulation of Fas-L induced by HIV-1 infection (see Figure 2D) strongly suggested that a Fas-mediated mechanism was involved in the viral-induced apoptosis of ex vivo CEM cells. Thus, we firstly assessed the FasL expression on ex vivo CEM cells at 48 h after virus challenge, as compared to uninfected ex vivo and parental CEM cells. The results clearly showed that HIV-1 infection induced a significant increase of FasL expression in ex vivo CEM cells using either SF162 (Figure 3B) or IIIB (data not shown) viral strains. To further explore the role of the Fas-pathway in the HIV-induced apoptosis in these cells, we treated ex vivo CEM cells with a blocking anti-Fas mAb contemporaneously to the HIV-1 infection and analyzed the percentage of apoptotic cells at 48 h. This set of experiments was performed exclusively with the SF162 M-tropic strain of HIV-1, in that the IIIB T-tropic strain progressively induced the well known cytopathic effect25 in both parental and ex vivo CEM cells. Thus, in order to avoid that virus-induced cytopathic effect may overlap the Fas-mediated apoptosis, we used SF162 as a virus prototype of Fas-induced apoptosis. The results showed that treatment with the blocking anti-Fas mAb did not significantly affect the background level of apoptosis in uninfected ex vivo CEM cells (Figure 3C, upper panel), while virtually abolished the increase of apoptosis induced by HIV-1 infection in ex vivo CEM cells (Figure 3C, lower panel). To further define the role of Fas/Fas-L in the HIV-induced apoptosis of ex vivo CEM cells, we performed in situ immunocytochemical studies. Firstly, the results confirmed that, at 48 h post-infection, the vast majority of ex vivo CEM cells were apoptotic (70–80%) (Figure 4A). Moreover, approximately 50–60% of apoptotic cells were Fas-L+ (Figure 4B) and the HIV-1-infected cells did not show apoptotic staining of the nuclei (Figure 4C). Furthermore, the apoptotic cells highly outnumbered the HIV-infected cells (70–80% versus 5–10%) (Table 1). Double staining experiments using flow cytometry or immunocytochemistry showed that the vast majority of Fas-L+ CEM cells were also Fas+ (data not shown).

Figure 4
figure 4

Immunocytochemical analysis of apoptosis, Fas-L expression and HIV-1 infection. (A) Apoptotic cells in cultures of HIV-1-infected ex vivo CEM cells (Cytospin preparation; Tunel Reaction+APAAP method, using NBT as chromogen) (×450 original magnification); (B) Double staining for Fas-L (orange) and Tunel Reaction (dark blue) in HIV-1-infected ex vivo CEM cells (Cytospin preparation; PAP versus APAAP methods respectively) (×850 original magnification), (C) Double staining for p24 (red) and Tunel reaction (dark blue). Tunel positive apoptotic bodies within and out the cells are shown (Cytospin preparation; APAAP method using fast red and NBT respectively as chromogen) (×850 original magnification)

Full size image
Table 1 Immunocytochemical staining for HIV-1 p24, apoptosis, Fas-L and Fas in ex vivo CEM cells infected with HIV-1, expressed as cell percentage±S.E.
Full size table

Low level chronic infection leads to resistance to spontaneous apoptosis

We then observed the behavior of uninfected and HIV-infected ex vivo CEM cells maintained in culture conditions. Of interest, the effects of the IIIB (T-tropic) strain of HIV-1 on ex vivo CEM cells highly differed from that of the SF162 (M-tropic) strain. In fact, the IIIB strain induced a massive cytopathic effect in ex vivo CEM cell cultures 8–10 days after in vitro infection, while the infection with SF162 HIV-1 strain did not induce in ex vivo CEM cell cultures cytopathic effect up to 30 days after infection (data not shown). Thus, we studied the apoptotic behavior of ex vivo CEM cells infected with the SF162 strain of HIV and maintained in culture conditions for 30 days, as compared to the ex vivo uninfected cells. The results firstly showed that ex vivo CEM cells infected with the SF162 strain of HIV-1 progressively lost their susceptibility to spontaneous apoptosis at 20–30 days post infection, while uninfected CEM cells maintained marked levels of spontaneous apoptosis (Figure 5). To further confirm these data we cloned SF162-infected ex vivo CEM cells, obtaining infected and uninfected clones. The results showed that all the SF162-infected clones exhibited levels of spontaneous apoptosis significantly lower than the uninfected clones (8–12% versus 19–25% respectively), suggesting that chronic infection with a M-tropic strain of HIV-1 induced an anti-apoptotic state in infected ex vivo CEM cells. To further explore this issue, we evaluated the expression of the anti-apoptotic complex Bcl-2/Bcl-xL/Bax in uninfected and SF162-infected ex vivo CEM cells. The RT–PCR analysis showed that there were no differences in the expression of Bcl-2, Bcl-xL and Bax mRNA between infected and uninfected ex vivo CEM cells and between infected and uninfected ex vivo CEM cells clones (data not shown). Flow cytometric analysis showed only a slight but significant increase in the percentage of Bcl-2+ cells in the SF162-infected ex vivo CEM cells, as compared to the uninfected cells (Figure 6A). However, the most impressive differences were detected by the in situ immunocytochemical analysis. In fact, the distribution of Bcl-2 highly differed between infected and uninfected cells. Ninety to 95% of SF162-infected cells showed an intense and diffuse cytoplasmic staining while virtually all the uninfected ex vivo CEM cells showed a highly localized pre-Golgi/perinuclear staining for Bcl-2 (Figure 6B, upper panel). Notably, the mitochondrial distribution did not differ between infected and uninfected cells (Figure 6B, lower panel). Both flow cytometry and immunocytochemistry showed that infected cells expressed Fas-L at lower level than uninfected cells at this time, while Fas expression did not differ between infected and uninfected cells (data not shown). Notably, either HIV-1-infected ex vivo CEM cells or clones were reinoculated s.c. in SCID mice resulting in the development of highly productive (500±150 pg/ml p24 serum levels) tumors within 20–30 days (data not shown), suggesting that in vivo or ex vivo selection of mutants with defective viruses did not occur in our experiments.

Figure 5
figure 5

Apoptosis in HIV-infected and uninfected ex vivo CEM cells maintained in cultures for 30 days. Percentages of apoptotic cells in uninfected ([open box]) and SF162 HIV-1 strain-infected ([hatched box]) ex vivo CEM cells at 1 day and at 30 days postinfection, analyzed by flow cytometry. Histograms are means±S.E.M. of five separate experiments

Full size image
Figure 6
figure 6

Bcl-2 expression and cellular distribution in HIV-infected and uninfected ex vivo CEM cells maintained in cultures for 30 days. (A) Flow cytometric analysis of the percentage of Bcl-2 positive cells in uninfected (dotted lines) and HIV-infected (solid lines) ex vivo CEM cells maintained in culture for 30 days; (‘B) Immunocytochemistry for Bcl-2 (upper panel) and mitochondria (lower panel) in HIV-1-infected (left panel) and uninfected (right panel) ex vivo CEM cells maintained in culture for 30 days. Arrows point to pre-golgi/perinuclear distribution of Bcl-2 in uninfected ex vivo CEM cells. Ninety to 95% of HIV-1-infected ex vivo CEM cells showed a more intense staining, widely distributed to the cytoplasm (left panel). (Cytospin preparation; APAAP method with fast red as chromogen+H&H counterstaining) (×850 original magnification). Results are representative of four separate experiments

Full size image

Discussion

The initial observation, which led to the ensemble of studies reported in this article, consisted in a marked delay in the growth of CEM s.c. tumors in SCID mice, after an acute HIV-1 infection. Further analysis showed that CEM cells obtained from uninfected s.c. tumors exhibited a certain level of spontaneous apoptosis and that HIV-1 infection markedly increased these baseline levels of apoptosis. We compared the phenotype of parental CEM cells (maintained in culture condition) to CEM cells obtained from SCID mice tumors (ex vivo CEM). The results showed that ex vivo CEM cells, differently to the parental cells, exhibited the memory phenotype (CD45RO+) and a sort of apoptotic phenotype characterized by: (i) spontaneous apoptosis; (ii) induced expression of Fas-L; (iii) reduced expression of Bcl-2. Both T-tropic and M-tropic strains of HIV-1 induced a marked increase of Fas-L expression and apoptosis in ex vivo CEM cells as early as 2 days post-infection. Moreover, the HIV-1-induced apoptosis was Fas-mediated and involved Fas/Fas-L+ uninfected cells. These findings strongly suggest that: (i) HIV-1 induces apoptosis of CD4+ T cells already prone to undergo apoptosis; (ii) HIV-1 induces a Fas-mediated autocrine fratricide/suicide of uninfected CD4+ T lymphocytes. Thus, some of our data appear to be consistent with recent reports showing that: (i) CD4+ T cells with low levels of Bcl-2 are highly susceptible to HIV-1-induced apoptosis; 26,27 (ii) HIV-1-induced apoptosis may be Fas-mediated28,29,30,31,32,33,34,35,36 and occurs predominantly in bystander cells.5,7,35 Notably, given the specific features of this model, the HIV-1-induced Fas-mediated apoptosis of uninfected cells occurs in a homotypic system of CD4+ T cells, suggesting that HIV-1 infection may trigger an autocrine suicide of CD4+ T lymphocytes without the participation of Fas-L expressing macrophages or CTL.37 The possibility that apoptosis can occur through a Fas/Fas-L-mediated autocrine suicide has been proposed in lymphocyte38,39,40 and in other cell systems.41,42 Here we show that the in vivo acquired pathway of spontaneous apoptosis is associated with the expression of the memory (CD45RO+) phenotype on CEM cells. This is consistent with some reports showing that CD4+ memory T cells are generally more susceptible to apoptosis than naive T cells.43,44 Even though the role of memory T cells in CD4+ T cell depletion is controversial,45 it is of interest to mention that some in vitro studies have shown that memory T cells are depleted by apoptosis more rapidly than naive T cells following HIV infection.11,46,47,48 Of interest, some body districts, such as intestinal mucosa, are particularly rich in CD4+ memory T cells and a consistent rate of Fas-mediated spontaneous apoptosis normally occurs in these districts.49 We have recently shown that human intestinal CD4+ T cells are naturally permissive to a long lasting HIV-1 infection, which is highly related to the memory phenotype, as well as to the expression of HIV-1 co-receptors and to a baseline state of activation.50 Notably, it has been reported that loss of CD4+ T cells in intestinal mucosa precedes the appearance of CD4+ T cell depletion in the peripheral blood of asymptomatic HIV infected patients.51 Taking into consideration all these findings, it is conceivable that activated memory CD4+ T cells are the preferential target for both HIV-1 infection and apoptosis-mediated CD4+ T lymphocyte depletion in AIDS patients.

In the present study, we have also explored the long-term effects of HIV-1 on ex vivo CEM cells by maintaining infected cells in culture for 20–30 days. The results have revealed a scenario of events dramatically different from that of the first 2 days post-infection. First, the T-tropic strain of HIV-1 induced a massive cytopathic effect in ex vivo CEM cells within the first 7–10 days of culture. The cytopathic effect induced by T-tropic strains of HIV-1 is a well known phenomenon due to the formation of enormous syncytia through a continuous virus-induced cell-to-cell-to-syncytia fusion.25 In contrast, the M-tropic strains of HIV did not induce any cytopathic effect, thus enabling us to investigate the effect of long lasting HIV-1 infection on apoptosis. Our results showed that: (i) the rate of spontaneous apoptosis progressively declined during time in culture, reaching values significantly lower than those of uninfected cells at 30 days of culture; (ii) the decrease of spontaneous apoptosis was associated with increased levels and diffuse cytoplasmic distribution of Bcl-2, as well as with lower levels of Fas-L, in infected ex vivo CEM cells as compared to the uninfected counterparts. It is well established that Bcl-2 protects cells from apoptosis.52 However, the precise subcellular localization of bcl2 and its importance in the modulation of programmed cell death are both unresolved problems. Particularly, bcl2 has been identified at the inner53 and the outer54,55 mitochondrial membrane, the endoplasmic reticulum,56,57 the nuclear membrane54,57 and close to the nuclei.58 Variation in the cellular distribution of bcl2 protein is associated with different susceptibility to apoptosis.58 Consistently with our data, resistance to apoptotic cell death is associated with a wide distribution of Bcl-2 within the cell, while the distribution of Bcl-2 to the perinuclear area is associated with an increased susceptibility to spontaneous apoptosis.58 The subcellular distribution of Bcl-2 is considered to be a crucial event in allowing interactions with the other apoptosis regulatory proteins (e.g. Bax). Thus, the wider distribution of Bcl-2 protein in HIV-1 infected ex vivo CEM cells could well explain their increasing resistance to spontaneous apoptosis.

Thus, these results appear to be consistent with the hypothesis that HIV-1 may in vivo prevent apoptosis through Bcl-2 mediated mechanisms until high levels of virus are produced. In this regard, it is worth mentioning that small amounts of some HIV-1-proteins have been shown to prevent or delay death of infected cells through various antiapoptotic mechanisms.2 As an example, low levels of endogenous expression of HIV-1 vpr can protect CD4+ T cells from apoptotic stimuli (including Fas pathway) through upregulation of Bcl-2.59 Moreover, our results suggest that a bcl2-mediated mechanism may render CD4+ T cells particularly susceptible to HIV-1-mediated CD4+ T cell death via a Fas pathway.60

In conclusion, our results, obtained in a model of SCID mice transplanted with CEM cells, are strongly reminiscent of a series of events possibly occurring in HIV-1 infected individuals. Particularly, our data indicate that an acute HIV-1 infection can induce a rapid CD4+ T lymphocyte depletion through a massive Fas-mediated autocrine suicide of differentiated memory CD4+ T cells, infecting a relative low number of lymphocytes. Following this dramatic scenario, CD4+ T lymphocytes infected with M-tropic strains of HIV-1, that do not induce cytopathic effect and predominate in the early phases of AIDS,61 undergo a chronic infection becoming resistant to apoptosis through the development of an efficient survival program. Notably, the dynamics of CD4+ T cell differentiation and cell phenotype, possibly shaped by the cytokine environment,62 may continuously change the CD4+ T cell permissiveness to HIV-1 strains which use CCR5 or CXCR4 as coreceptors. Thus, on one hand, HIV-1 may induce a massive Fas-mediated autocrine suicide/fratricide of uninfected CD4+ memory T cells during acute infection. On the other hand, HIV may inhibit or at least delay spontaneous apoptosis in chronically infected CD4+ T cells allowing a long lasting viral production. These cells may constitute a pool of latently infected CD4+ T cells that are rapidly established during primary infection23 and are resistant to highly active antiretroviral therapy.24

Lastly, these data emphasize the usefulness and versatility of human-SCID mouse models in investigating in vivo the relationships between T cell differentiation, HIV-1 infection and the mechanisms underlying CD4+ T cell depletion.

Materials and Methods

Growth of CEM cell-tumors in SCID mice and in vivo HIV-1 infection

CB.17 SCID/SCID female mice (Charles River, Milan, Italy) were used at 4–5 weeks of age and were kept under specific pathogen-free conditions. SCID mice were housed in microisolator cages and all food, water and bedding were autoclaved prior to use. The animal studies were performed in biosafety level three facility. Mice were injected subcutaneously (s.c.) in the shoulder with 20×106uninfected CEM cells63 resuspended in 0.2 ml RPMI 1640 medium. The two major diameters of each tumor nodule were measured by callipers and the mean tumor diameter was calculated for each tumor. To deplete animals of some residual reactivity, SCID mice were injected i.p. with 0.2 ml of a monoclonal anti-mouse granulocyte antibody (RB6-8C5 hybridoma), 1 day before, 3 and 7 days after cell injection, as previously described.63

The in vivo HIV infection of CEM-SCID mice was performed with a simultaneous s.c. injection of 20×106 uninfected CEM cells with 106 TCID50 cell-free virus. The viral strains used in these experiments were: HIV-1IIIB and HIV-1SF162. In all conditions, the HIV-infected chimeras were sacrificed when the tumors reached 20–25 mm mean diameter and analyzed for the virus replication at the tumor site and p24 antigenemia.

Detection of viral infection

At sacrifice, the CEM cell tumors were excised. To obtain single cell suspension tumors were carefully minced, mechanically disrupted with the blunt end of a 5-ml syringe plug and filtered. The single cell suspensions were washed twice in RPMI 1640 medium. Cell suspensions underwent: (i) HIV-1 DNA-PCR as previously described;19 (ii) HIV-1 RT–PCR, using specific primers which detect all viral RNAs, as reported elsewhere;64 (iii) FACS analysis for human cells markers and apoptosis. Sera of infected animals were tested for HIV p24 antigen by an antigen capture enzyme-linked immunosorbent assay (Dupont, B-1130 Bruxelles, Belgium).

Cell culture and in vitro HIV-1 infection

Parental CEM cells (ATCC, Rockville, MD, USA) as well as CEM cells collected from s.c. tumors of SCID mice (ex vivo CEM cells) were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum. Cells were seeded at 2×105/ml and passaged every 3 days. For in vitro HIV-1 infection cells were pelleted and incubated with the virus inoculum at 0.1 M.O.I. for 1 h at 37°C, washed three times and cultured in complete medium. The virus stocks were derived from clarified culture medium of PHA-stimulated HIV-1-infected PBMC, frozen at −140°C. Titers were determined by standard end-point dilution methods. The virus strains used in these experiments were: HIV-1IIIB, HIV-1SF162.

Cell cloning

CEM cells obtained from subcutaneous tumors were exposed to HIV-1SF162 virus strain at 1 M.O.I. for 1 h at 37°C, then washed extensively, and seeded in complete medium at a concentration of 0.5 cell/well in a 96-microwell cell culture plate. After 2 weeks pure clones were screened for virus infection by PCR. Infected as well as uninfected clones were propagated for further analyses as described.65

Western blot analysis

Cell extracts were resuspended in SDS sample buffer, denaturated by boiling and separate on 8% SDS–PAGE gels. Then, proteins were transferred to Hybond C Extra (Amersham Pharmacia Biotech, NY, USA) and blocked in 5% milk overnight. Bcl-2 was detected with an anti-Bcl-2 monoclonal antibody (Santa Cruz, CA, USA) and visualized with peroxidase anti-Ig followed by ECL (Pierce, SuperSignal Substrate, Rockford, IL, USA). Immunoblotting for actin protein (Chemicon, CA, USA) was performed to normalize the level of proteins in all the samples.

Flow cytometry and apoptosis detection

CEM cells were washed twice with phosphate-buffered saline and stained for 30 min with the appropriate monoclonal antibody at 4°C. The monoclonal antibodies used were anti-CD45RA FITC, anti-CD45RO PE (Becton Dickinson, Mountain View, San Diego, CA, USA), anti-FAS PE (Becton Dickinson, San Diego, CA, USA), anti-FAS-L (NOK-1, Pharmingen, CA, USA), anti-Bcl-2 (Neomarkers, Fremont, CA, USA). An anti-mouse IgG mAb PE (SIGMA) was used for indirect staining. Seventy per cent ethanol treatment was used to permeabilize cells before staining for Bcl-2. After staining cells were fixed with 2% paraformaldehyde and analyzed on a FACSORT cytometer (Becton Dickinson) equipped with a 488 nm argon laser. Data were recorded and analyzed by using LYSIS II software (Becton Dickinson). The analysis was performed by gating ex vivo CEM cells using forward and side scatter characteristics, to avoid SCID mouse cell contamination.

Apoptotic cells were detected by labeling DNA strand breaks with FITC-dUTP by terminal transferase (TUNEL reaction) (‘In situ cell death detection kit’, Boehringer Mannheim). Briefly 106 cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, 0.1% sodium cytrate and labeled at 37°C for 60 min with TUNEL reaction mixture as recommended by the manufacturer. Apoptosis was evaluated directly by flow cytometry.

Fas blocking

Ex vivo CEM cells were infected with HIV-1IIIB or HIV-1SF162 strains in the absence or in the presence of an anti-Fas ZB4 neutralizing mAb (10 μg/ml) (UBI, NY, USA). Apoptosis was evaluated by FACS analysis 48 h postinfection.

RT–PCR for apoptotic regulating genes

mRNA for human pro-apoptotic and anti-apoptotic genes (Fas, Fas-L, Bcl-2, Bax, Bcl-xL) were detected in CEM cells by RT–PCR as previously described19 with specific primer pairs: Fas-L 5′ CAGCTCTTCCACCTACAGAAGGAG, 3′ CAGAGAGAGCTCAGATACGTTGAC; primers specific for Fas, Bcl-2, Bax, Bcl-xL are reported elsewhere.66,67,68,69 The samples were amplified for 30–35 cycles at the following conditions: 94°C 40 s, 62°C 40 s, 72° 40 s. β2-microglobulin RT–PCR19 was run in parallel to normalize the levels of human RNA in all the samples.

Immunocytochemistry

Cell suspensions from the tumors were spun onto glass slides (Shandon, Cheshire, UK) or attached to L-polylysine-covered glass chamber slides (Labtek Naperville, IL, USA),70 and stained by immunocytochemistry for HIV-p55/p18 (anti-HIV-p55/p18 IgG1, clone 11H9, from Medical Research Council AIDS Reagent Project, London, UK), Fas (Pierce, Rockford, IL, USA), Fas-L (Pharmigen, San Diego, CA, USA), Bcl-2 (Biomeda, Foster City, CA, USA), mitochondria (Chemicon, CA, USA) and TUNEL reaction (‘In situ cell death detection kit′ Boehringer Mannheim),71 using the alkaline phosphatase anti-alkaline phosphatase (APAAP) (Dako, Denmark) method or the peroxidase-anti-peroxidase (PAP) (Dako, Denmark) method, in single and double staining, as appropriate.70